Si/Ge Feedback Loop Recovery Stage Question

[lars-musik]

Because I am still looking for THE overdrive for my setup, I have been looking into the King of Tone (or better: 1/2 of it, a Prince of Tone) and played around with the clipping options. I came up with a pedal with some switching options (silicon and germanium, symmetrical and aysmmetrical in the feedback loop plus mosfet/silicon in the shunt-to-ground option):


 Because of the volume drop when switching to germanium I placed 8 of these diodes (Vf ~ 0,255 V) in the feedback loop, but still the volume drop is significant (as you might expect, 'cause the Vf-ratio of si:ge is still approx. 1.5:1.0).

So, I would like to have a recovery volume stage for the germanium loop, at best adjustable, so I could experiment with fewer germanium diodes there (the difference between 8 germanium and 4 silicon is minimal to my ears, so maybe fewer germanium diodes would gain more audible clipping?).

I searched for "recovery gain stages" and the like and found that these are quite common (like in the Big Muff after the tone control and in other designs) but due to a lack of electronic knowledge I am unable to come up with a circuitry that is likely to work.

Somebody willing to show me the way?

Thanks, Lars

[antonis]

I think that it would be more convenient to "drop" the Si loop output (by a series resistor) to the same level as the Ge one and then decide if you need more volume..

And if so, the gain recovery will be the same for ALL your choices..

[lars-musik]

Thanks for your suggestion, maybe it is a good idea. However, there's a minor detail missing in the schematic. The toggle switch between Si and Ge is a On-Off-On switch and in the "off" position the pedal acts as a boost. A lower Si-clipping volume wouldn't be my my first choice, because then the boost (which is about unity gain to the Si clipping now) would then stand out volumewise.

Maybe it's an amateurish thought: but isn't there a possibility to add some kind of a boost right after the germanium diode cluster and before the switch?

[antonis]

I'm not sure if I've got you right but the "pure" boost will remain unaffected...

You can allways add a common emitter gain recovery stage (like BMPi) but you'll have to put a trimmer in place of Rc or Re for experimenting with various number (and type) of clipping diodes..
(althought I think that it should be better placed at the end of the circuit for final recovering - including tonestuck's loss)

[lars-musik]

Could this work?

 To my excuse: I just got myself my first breadboard and haven't set it up yet. So all of this is going to be tested on my already soldered pcb (along with the breadboard I got myself about 100m of desoldering wick...). So I'd like your approval before ruining the board.
Thanks!

[antonis]

I've a VERY slow PC so I can't draw, load and upload photos, sorry..

Your BJT needs bias so you have to connect on it's base a voltage divider from Vcc to GND (say 470k/150k) - or even simpler, a high value resistor (510k or so..) from VB to base...

And there isn't need for 2 trimmers - just a 25k multi-turn at the collector with an emitter resistor of, say, 3k3...


Ooops, and a coupling capasitor of, say again, up to 10μF from the collector to U2...
(the exact value should be calculated taking in mind the tonestuck's impedance but I think that if you ommit it shouldn't cause major problem...)

[lars-musik]

Thank you very much antonis,

might this do?


However, this feels very much like painting by numbers (and I'd like to understand a little more, if possible). I can see how collector and emitter resistance varies the output level. But how is the biasing voltage divider calculated? And how does the capacitor from collector to U2 influence the clipping cluster? And finally (something I never really grasp): I do influence that strange thing "impedance" with the circuit, don't I? Good or bad?

Thanks again!

[digi2t]

How about this one? It's the final gain stage of the Gemini III Dual Fuzz Generator. Jimi used it in the Suzi Q (I think), and said it worked great.

[duck_arse]

lars, your new diagram shows a short drawn across C13. this will not do. and, seeing as there are set DC conditions around that opamp, you will also need a dc blocking cap between the diode strings and the bias resistors on the transistor base. (I'm sure antonis has also typed this, but his pc is so sloooowwwww it hasn't arrived yet.)

for setting up a single stage transistor amplifier, [lord sir] gibsonGM re-showed a thumbnail method devised by PRR in a thread recently. the name of the thread is beyond me at the moment, search (or maybe risk bothering gibson).

I'm pretty sure yr boost stage is not the right method, though. it should (should) be easy enough to slip in a few extra poles on your switches to select a resistor divider for the higher outputs. did you say if you had a breadboard?

I'm so slow of typing that digi2 got in before me.

[Gus]

something I posted here before look at the last stage


A link
http://www.gmarts.org/index.php?go=217
In the linked circuit switch the last stage gain when Ge is selected

[lars-musik]

Allright.  Here's the new attempt according to digi2t's post. Any objections? Hope this works. Maybe tomorrow I'll find the time to inaugurate my breadboard.



 

[PRR]

[lars-musik]

Thanks very much Paul. I know your solutions usually turn out to be great. Would to mind to elaborate? To which of the before mentioned circuits do you refer? I  Or am I missing something and this is something completely different?

[PRR]

> To which of the before mentioned circuits do you refer?

This is just the diode-clip stage of the plan you posted, with trim-pot and resistor added. It shifts-up the gain in the "Ge" side by a trimmable amount to match the "Si" side.

[lars-musik]

With an added trimpot and resistor to ground from the Ge-clippnig section I can image how I can lower the volume – not rise it? I just tried it (I cut the trace between the D11/D15 – U2 junction and inserted the 1K resistor to ground and a 1K pot) and the thing just falls silent. Have I got this right: D11/D15 to the middle lug of the pot, one lug to the switch, one lug to ground?

Before Paul's answer came up I finally got to breadboarding a recovery volume stage. But the sound was very dissapointing : Yes it is louder but the Ge-diodes sound dull and the stage introduces a kind of white noise into the whole circuit I did not wish for (maybe due to the long cables to the breadboard?).

 I have to admit, all this is not worth the hassle. I can dial in the sound I want albeit quieter on the Ge-side so I guess I'd just box it up (still enough work to do). Anyway, this is the last  version and a picture of my first breadboarded circuit (!!)



Thanks for your support!

[PRR]

> I can image how I can lower the volume – not rise it?

This is *inside* the feedback loop. It lowers the signal to the diodes, not to the output. If the resistor divider is set to 50%, and the diodes clip at 1V, it takes 2V out of the opamp to make that happen.

Your "last version" seem to work on some totally different concept. I suspect it clips in the (totally Silicon) 2N5088, not in the Germanium diodes.

[lars-musik]

So I finally boxed the Professor up, because I really like this one, even with the difference in volume. This is my first build with onboard jacks, switches and pots. The flip switches are drowned it hot glue to make them stable enough for operation without washers (they do not fit due to the height of the 3PDT).

[duck_arse]

looks good, lars, glad you got it "finished". I can't see your pots, are they 9mm or 16mm? if you wanna hide your nuts, switchwise, you might try using a sub-panel. done properly (ie, not the way I do it, where something allays goes wronk), you can mount the pots and switches to the same panel as the pcb, and the panel can be any old scrap you have, even bent to fit if you like. then you have no pot nuts on the front to mess paint, and you can drill/size the switch holes so the switch barrel is a nice close fit.

how you mount the panel to the box can be tricky. you can use the footswitch at one end (you can even mount the footswitch to the panel if you have enuff hardwares), and maybe hang a tab off the other end to mate with the dc socket.

[lars-musik]

The pots are 9mm to which I soldered leftovers of resistor legs to make them stand tall enough. The pcb is held in place now by the 3pdt and the jacks and the pots.  The switches are pretty solid soldered and glued on there. The biggest problem with that build is that I can't really hear a difference between all these switching options. So I think, I'll just figure out what setting I like best a build another one in a 1590a.
 Then I might consider hiding my nuts!